def alert_for_schema_change(db, check, conf):
df = get_last_results(db,
check,
Metric.TABLE_METRIC,
Metric.SCHEMA_CHANGE,
conf,
days=1)
for index, row in df.iterrows():
changes = json.loads(row[0])
if changes["operation"] == "table detected":
continue
alert = Alert(
text=f"""
schema change detected - {changes['operation']}: {changes['column_name']}
""",
severity=2,
table_id=check.table_id,
alert_type=check.name,
created_at=conf.for_time,
)
metrics_session.add(alert)
metrics_session.commit()
def volume_alert(db, table, conf):
sql_df = get_last_results(db, table, 'metrics_data_volume', conf)
filtered = sql_df[sql_df['time_interval'] == settings.INTERVAL_FOR_ALERTS]
checked_txt = f'volume in interval: {settings.INTERVAL_FOR_ALERTS}'
alert_on_z_score(filtered, table, 'count', 'volume_alert', checked_txt,
conf)
def delay_alert(db):
tables = MonitoredTable.get_monitored_tables(db.name)
for table in tables:
sql_df = get_last_results(db, table, 'metrics_data_delay')
checked_txt = f'delay since last data'
alert_on_z_score(sql_df, table, 'value', 'delay_alert', checked_txt)
def volume_alert(db):
tables = MonitoredTable.get_monitored_tables(db.name)
for table in tables:
sql_df = get_last_results(db, table, 'metrics_data_volume')
for interval in settings.VOLUME_INTERVAL:
filtered = sql_df[sql_df['time_interval'] == interval]
checked_txt = f'volume in interval: {interval}'
alert_on_z_score(filtered, table, 'count', 'volume_alert',
checked_txt)
def alert(db, check, conf):
if check.name == Metric.SCHEMA_CHANGE:
alert_for_schema_change(db, check, conf)
else:
metrics = check.metrics
for column, metrics_for_col in metrics.items():
for metric in metrics_for_col:
sql_df = get_last_results(db, check, column, metric, conf)
sql_df["result"] = pd.to_numeric(sql_df["result"])
alert = column + "_" + metric
checked_txt = alert + " is failing"
alert_on_z_score(sql_df, check, alert, checked_txt, conf)
def alert(db, check, conf):
if check.name == Metric.SCHEMA_CHANGE:
alert_for_schema_change(db, check, conf)
else:
metrics = check.metrics
for column, metrics_for_col in metrics.items():
for metric in metrics_for_col:
sql_df = get_last_results(db, check, column, metric, conf)
sql_df["result"] = pd.to_numeric(sql_df["result"])
if column == Metric.TABLE_METRIC:
alert = metric
else:
alert = name_for(column, metric)
checked_txt = alert
alert_on_z_score(sql_df, check, alert, checked_txt, conf)
def values_alert(db):
tables = MonitoredTable.get_monitored_tables(db.name)
for table in tables:
sql_df = get_last_results(db, table, 'metrics_data_values')
checks_df = sql_df[['check_name', 'time_interval',
'column_name']].drop_duplicates()
for i, row in checks_df.iterrows():
df = sql_df[(sql_df['check_name'] == row['check_name'])
& (sql_df['time_interval'] == row['time_interval']) &
(sql_df['column_name'] == row['column_name'])]
check_text = f'values for {row.check_name} in column: {row.column_name}, interval: {row.time_interval}'
alert_on_z_score(df, table, 'check_value', row['check_name'],
check_text)
def values_alert(db, table, conf):
sql_df = get_last_results(db, table, 'metrics_data_values', conf)
sql_df = sql_df[sql_df['time_interval'] == settings.INTERVAL_FOR_ALERTS]
checks_df = sql_df[['check_name', 'column_name',
'column_value']].drop_duplicates()
for i, row in checks_df.iterrows():
df = sql_df[(sql_df['check_name'] == row['check_name'])
& (sql_df['column_name'] == row['column_name'])]
if row['column_value']:
df = df[df['column_value'] == row['column_value']]
check_text = f'column: {row.column_name}, values for {row.check_name}'
alert_on_z_score(df, table, 'check_value', row['check_name'],
check_text, conf)
def delay_alert(db, table, conf):
sql_df = get_last_results(db, table, 'metrics_data_delay', conf)
checked_txt = f'delay since last data'
alert_on_z_score(sql_df, table, 'value', 'delay_alert', checked_txt, conf)